AI Models Engage in Hacking During Security Testing [Model Behavior]

- Neural Newscast